AI is Solow Paradox at Work

An analysis of 4,500 work-related artificial intelligence use cases suggests we are only in the very-early stages of applying AI at work and that most of the use cases have not yet moved to a stage where we can measure return on investment or productivity impact. 

That is worth keeping in mind. 

Most use cases so far only affect speed or time savings. Few use cases are more-directly integrated into customer-facing revenue-generating activities. 

The vast majority of use cases are very basic, says a Section AI report. Some 14 percent of workers say their most valuable AI use case is Google search replacement. As helpful as that might be, it is hard to measure productivity gains at this point. 

source: Section AI

About 17 percent of workers use AI for drafting, editing, and summarizing documents. Again, productivity improvements are difficult in those cases, but perhaps more measurable in terms of time savings. 

So far, Section AI researchers found only two percent of users have built automations for copy generation, which would save more time, for example. 

About three percent say their most valuable use case is data analysis or code generation, and there the ROI seems easiest to document in terms of time saved or effort avoided, rather than other revenue-generating metrics. 

source: Section AI

In fact, nearly a quarter of respondents say AI does not save them any time at all, which might seem odd unless those users are having to spend time learning how to use AI, which would, in fact, take more time. 

In other cases, they might find they are having to spend time checking the answers and output, which again might take additional time. 

The point is that we are in early stages of deployment, where it remains difficult to assess productivity gains. 

source: Section AI

As unhelpful as it might be, transformative technologies often fail to show up in productivity statistics for years, or even decades, after their introduction, as the Solow Paradox describes. 

Measuring language model impact by "minutes saved per task" captures only the shallowest layer of value, many would argue. The reason is that what we can measure sometimes is not all that important. 

Productivity metrics are generally designed to measure output per hour (quantity). They are notoriously bad at measuring quality. 

If a model helps a software engineer write safer, more robust code, or helps a marketer generate a campaign that resonates better with customers, standard productivity metrics might show zero gain (or even a loss.

Also, In the early stages of adoption, productivity often dips, since firms and workers must invest time and capital into training, restructuring workflows, and figuring out how to use the new tools. 

This "intangible capital" investment does not produce immediate revenue.

Also, as always, adopters are using language models to do existing tasks faster (writing emails). True productivity explosions only occur when businesses re-architect their entire workflows to do things that were previously impossible, rather than just speeding up legacy processes.

Innovation	Initial Era	The "Lag Phase"	Primary Reason for Lag	When Productivity Finally Spiked
Electric Power	Late 1880s (Electric motor introduced)	~30–40 Years	Factory owners swapped steam engines for electric motors without changing factory layouts.	1920s: When "unit drive" systems allowed for the assembly line and decentralized manufacturing.
Computers (IT)	1970s–80s (Mainframes & PCs)	~15–20 Years	The "Solow Paradox." Computers were used for isolated tasks (word processing) rather than networked data flow.	Mid-1990s: When the internet and enterprise software (ERP) enabled supply chain integration and instant communication.
The Internet	Early 1990s (World Wide Web)	~10–15 Years	The "Dot Com Bubble." Investment rushed in, but business models (e-commerce, cloud) were immature.	Late 2000s/2010s: When mobile internet, cloud computing, and smartphone adoption created the app economy.
Generative AI (language models)	2022–Present (ChatGPT moment)	Ongoing	Current focus is on "task replacement" (writing, coding) rather than "workflow redesign" (autonomous agents, new R&D methods).	Prediction (2027–2030+): Likely when AI moves from a "copilot" (assistant) to an "agent" that can autonomously execute complex, multi-step workflows.

That sort of measurable productivity gain cannot be demonstrated so soon. 

Can Netflix Become Disney Faster than Disney Can Become Netflix?

To a larger degree than might be immediately obvious, the new Netflix challenge might be whether “Netflix can become Disney faster than Disney can become Netflix.”


Source: Nano Banana generated image

It might appear that the prior challenge (since the 2013 to 2015 period), for Netflix to "become HBO faster than HBO becomes Netflix," has mostly been achieved.

That statement encapsulated the business challenge of Netflix becoming a content powerhouse faster than HBO could become a streaming giant.

Netflix largely succeeded, and an acquisition of Warner Brothers Discovery, though a gamble, might help with the next challenge, which is for Netflix to diversify and broaden its revenue sources to maintain higher growth rates as the core streaming business slows.

And that is where the Disney challenge is relevant, as many expect future Netflix growth to rely on new sources such as monetization of intellectual property (merchandising, for example); theme park or other experiential revenue streams; while maintaining a successful original content creation and distribution mechanism.

"Total entertainment share" now becomes a relevant objective for Netflix, which can seek greater wallet share across advertising revenue; live sports programming; merchandise; gaming; events and bundles of products.

As it once was said that “Disney has the mouse,” so Netflix might hope to expand from content creation and distribution to leverage its intellectual property in various other venues, ranging from events and physical entertainment attractions to live sports and gaming.

Netflix revenue growth and profit margins might hinge on success in that effort.

Logs and Splinters

"Why do you see the speck in your neighbor's eye, but do not notice the log in your own eye? Or how can you say to your neighbor, 'Let me take the speck out of your eye' while the log is in your own eye? You hypocrite, first take the log out of your own eye, and then you will see clearly to take the speck out of your neighbor's eye." (Matthew 7:3-5).

That passage is a classic (and exaggerated for effect) reminder that we humans are hypocrites.  A reasonable “definition” is that “a hypocrite is someone who believes they have virtues, beliefs, or feelings they don't actually possess. 

Many understandings might emphasize “pretending to have certain virtues,” but the more-common application might simply be that we often fail to live up to our own best versions of ourselves. We fall short, even when aiming at the target. 

It is less an epithet (“you’re a bad person) and more an observation about “human nature.” 

It might apply when a person’s actions contradict their stated ethical standards. The word is from the Greek word for “stage actor,” which gives you the flavor. 

More generally, it simply refers to the failings we humans have when judging others without recognition of our own failings. Consider the oft-heard phrase about “threats to democracy” or the claim of efforts to “protect democracy.”

Lots of logs, we might also say. 

Without agreeing in any way with some “executive power” embellishments President Trump might arguably prefer; his personality traits or anything else people may dislike about him, the frequently-heard and heated “threat to democracy” litany is disingenuous and betrays a stunning lack of objectivity. 

Or just call it hypocrisy. 

One might just as well allege that the actions taken by many who say they are Democrats pose equivalent or greater “threats” to democracy. 

Presumably because they see the president as an “existential threat,” almost any excessive remedies can be proposed, including violence of many sorts that overrides our settled governance procedures (elections, non-partisan judicial review, rule of law, “letting the system work”). 

But we also have seen other actions, ranging from efforts to disqualify the president as a candidate; politically-motivated impeachments; sentiments to “pack the Supreme Court;” actual suppression of some views; spreading action misinformation; efforts to deplatform opponents and bias in the mainstream media’s reporting. 

To be fair, many well-intentioned people will argue their proposals are not extreme; not dangerous to democracy; not extra-legal or inappropriate. In practice, language and action might be just as much an objective  “threat to democracy” as anything they believe the president is doing. 

AActor	Action	Description	Perceived Threat to Democracy
DDemocrat Politicians	Use of violent or elimination rhetoric against political opponents	Examples include President Biden referring to "putting Trump in a bullseye" and Rep. Dan Goldman stating Trump "has to be eliminated" (later apologized). Other Democrats have used similar language portraying Trump as an existential threat.	This rhetoric is seen by critics as inciting violence or normalizing extreme measures against opponents, potentially eroding civil discourse and leading to political violence.
DDemocrat Politicians	Efforts to disqualify political opponents from ballots	Democratic officials in states like Colorado and Maine attempted to remove Trump from primary ballots citing the 14th Amendment's insurrection clause.	Viewed as subverting voter choice and using legal mechanisms to eliminate competition, bypassing the electoral process.
DDemocrat Politicians	Partisan impeachments and investigations	Democrats impeached Trump twice, which Republicans argue were politically motivated without sufficient evidence.	Perceived as weaponizing congressional powers for partisan gain, undermining the legitimacy of oversight and eroding institutional norms.
DDemocrat Politicians	Proposals to alter democratic institutions	Suggestions to pack the Supreme Court, abolish the filibuster, or eliminate the Electoral College.	Critics argue these changes would concentrate power in the hands of the majority party, threatening minority rights and balanced governance.
DDemocrat Politicians	Support for censorship or restrictions on speech	Some Democrats have advocated for social media bans on certain groups (e.g., a Georgia Democrat candidate proposing a 4-year social media ban for MAGA voters).	Seen as infringing on free speech rights, a cornerstone of democracy, to silence opposition.
DDemocrat-Supporting Media	Spreading misinformation or biased reporting	Liberal media outlets accused of amplifying false narratives, such as the Russia collusion story, or suppressing stories like the Hunter Biden laptop.	Erodes public trust in institutions, polarizes society, and misinforms voters, undermining informed democratic participation.
DDemocrat-Supporting Media	Inciting division through partisan coverage	Networks like MSNBC portray Republicans as threats to democracy, using inflammatory language.	Contributes to political polarization and hostility, potentially leading to violence and weakening democratic cohesion.
DDemocrat-Supporting Media	Collaborating with tech platforms for censorship	Pressure on social media companies to deplatform conservative voices or label content as misinformation.	Limits free expression and access to diverse viewpoints, creating echo chambers and biasing public discourse.

No such criticisms excuse fair assessment of language, tone or action on the part of the president that some find offensive. But neither is it a fair assessment that the main or only threats to “democracy” come only from one side. 

They come from everywhere, from all of us, all the time.

AI Impact: Analogous to Digital and Internet Transformations Before It

For some of us, predictions about the impact of artificial intelligence are remarkably consistent with sentiments around the importance of innovation in the technology business we have heard since the 1980s and up through the internet revolution.

“In 2030, I think we will be bringing offerings and solutions to market that we can’t even envision

today because the technology isn’t there yet,” says Maureen Power Sweeny, RapidScale chief revenue officer.” I would say 50 percent of our revenues will come from new offerings.”

Some of us have been hearing that since the 1980s, and the sentiment seems borne out by events, and especially for digital products that can be created and adopted faster than older physical-only products. 

Five to 10 years is now multiple generations of product evolution in software, cloud, data, and AI.

Company	Product generating major revenue today	Year introduced (approx.)	What it displaced or superseded
Amazon (AWS)	Lambda, SageMaker, Bedrock	2014–2023	EC2-only compute, on-prem ML
Microsoft	Azure OpenAI, M365 E5, Copilot	2019–2024	Perpetual Office licenses, on-prem Exchange
Salesforce	Einstein AI, Platform, Industry Clouds	2016–2023	Standalone CRM licenses
NVIDIA	Data center AI accelerators	2017–present	Gaming GPUs as primary revenue driver
Adobe	Creative Cloud subscriptions	2013–2017 ramp	Perpetual boxed software
ServiceNow	Creator workflows, AI ops	2018–2024	ITSM-only positioning
Zoom	Phone, Contact Center, AI Companion	2019–2024	Single-product video meetings
Apple	Services (TV+, Pay, Arcade, Fitness)	2019–present	Hardware-only growth
Google	Cloud AI APIs, Workspace AI	2018–2024	Ad-only monetization model
Netflix	Ad-supported tier, originals	2016–2023	Licensed content distribution

But many industries feature high rates of innovation and product replacement. In many fast-moving consumer categories, 40 percent to 70 percent of active SKUs (stock keeping units) at major retailers are replaced within five years, even though the brand name on the shelf barely changes.

Category	Brand	Current high-volume SKUs (examples)	Introduced (approx.)	What they replaced or supplemented
Beverages	Coca-Cola	Zero Sugar reformulations, mini-cans	2017–2023	Full-sugar flagship SKUs
Snacks	PepsiCo (Lay’s)	Baked, Kettle, limited flavors	2016–2024	Core salted potato chips
Dairy alternatives	Danone	Oat-based yogurts, creamers	2018–2024	Traditional dairy SKUs
Household	P&G (Tide)	Pods, hygienic clean, cold-water	2015–2023	Liquid detergent bottles
Beauty	L’Oréal	Clean beauty, dermocosmetic lines	2017–2024	Mass cosmetic formulations
Food	Nestlé	Plant-based frozen meals	2019–2024	Animal-protein frozen SKUs
Alcohol	AB InBev	Hard seltzers, flavored malt drinks	2018–2023	Core beer SKUs
Personal care	Unilever	Sulfate-free, refill formats	2019–2024	Traditional shampoo bottles
Candy	Mars	Sugar-reduced, portion-controlled	2016–2023	Standard single-serve bars
Pet food	General Mills	Fresh, functional pet foods	2019–2024	Dry kibble-only lines

So one commonality with prior industry changes is rapid and continual product innovation. 

The other perhaps-obvious potential change is industry disruption. 

In 2030, enterprise success won’t be measured by steady progress toward long-term targets but by how much an enterprise disrupts its industry quarter by quarter, IBM researchers argue. 

The biggest risk won’t be making the wrong bets, but making bets that are too small, the authors argue. 

That again will sound familiar to those who witnessed the changes in firms and industries wrought by the internet. Netflix provides an example. 

Netflix disrupted multiple stages of the video entertainment value chain, first by unbundling physical retail rentals and later by collapsing production, aggregation, and direct distribution into a single global streaming platform.

source: Digital Leadership 

Value chain stage	Traditional role	Netflix’s move	Why it was disruptive
Content creation & commissioning	Studios and networks funded, developed, and greenlit shows and films, often for domestic first windows.	Became a major commissioner and producer of originals, including local-language series with global ambitions.academic.oup+1	Shifted power from broadcasters to a data-driven buyer with global reach, reducing reliance on legacy networks and altering bargaining dynamics.academic.oup​
Aggregation & programming	Broadcasters and cable networks assembled schedules and channel line‑ups; video stores curated shelves.	Offered a huge on-demand catalog with personalized rows and recommendations instead of fixed schedules or shelves.accio+1	Undermined the value of linear programming and physical curation, making algorithmic discovery the primary gatekeeper of attention.accio+1
Distribution to consumers	Cable/satellite operators and retail chains controlled access via physical locations or set‑top boxes; households subscribed to bundles or rented per title.forbes+1	Delivered content over the open internet to apps on TVs, phones, and PCs via a flat-rate subscription.accio+1	Bypassed cable operators and retail stores, drove cord‑cutting, and shifted consumer spend from channel bundles and rentals to streaming subscriptions.intrinsicinvesting​
Retail / rental monetization	Video rental stores monetized individual rentals and late fees, constrained by local inventory and store hours.linkedin+1	Introduced queue-based subscription rentals by mail, then streaming with no late fees and no trip to the store.linkedin+1	Eliminated the core revenue levers of incumbents (late fees, impulse rentals), making their store-heavy model uncompetitive.linkedin+1
Marketing & discovery	Studios and networks relied on mass advertising, trailers, and prominent placement in stores or schedules.	Used in‑product recommendations, personalized homepages, and data-informed promotion of titles.accio+1	Reduced dependence on external marketing funnels, internalized discovery, and extended tail consumption of niche titles.accio​
International sales & licensing	Rights sold territory-by-territory with complex windowing across cinema, pay TV, free TV, and home video.	Negotiated multi-territory and global rights and launched near-simultaneous international releases on one platform.academic.oup+1	Collapsed layers of local intermediaries and windows, standardizing global access and weakening traditional territorial carve‑ups.academic.oup​

The internet accelerated innovation by making information, markets, and coordination vastly cheaper and faster, turning experimentation by firms into a continuous, global, software-driven process. 

Enablers included:

• Radically lower information and transaction costs 

• Continuous software-based iteration, shortening innovation cycles from years to weeks or days.

• Global market access​

• Real-time feedback loops

• Platform and ecosystem effects (e-commerce, app stores, social networks).

The take-away is that, as we have seen before, AI is going to enable more industry and firm disruption.